In this study, we have successfully fabricated a collagen/PCL composite fibrous material by electrospinning and explored its application as nerve guide substrate or conduit in vitro and in vivo.
As a simple and versatile fabrication technique, electrospinning is a good choice to make fibrous scaffolds such as nerve conduits with tailored porosity and degradation rate, which can easily provide mass production of different-sized conduits made of nanofibers or submicron fibers [2, 14]. So far diameter and morphology of electrospun fibers are controlled mostly empirically, relying on variation of solution concentration, molecular weight of the polymer and ratios of different polymers for composite material electrospinning . Based on these principles and our previous explorations, we tried parameters of total solution concentration (10%) and collagen/PCL weight ratio (1:1) in this study, and finally successfully produced fibrous meshes and porous nerve conduits with fibers of 675 ± 386 nm in diameter.
As is known, PCL is one kind of aliphatic and degradable polyesters with good tensile properties and slow degradation kinetics. But its bioinert and hydrophobic properties limit its use as a biomedical material . Collagen as the main structural protein of ECM possesses excellent biocompatibility. Yet as a biomedical material for tissue engineering, it often requires additional chemical treatments of cross-linking to decrease its degradation rate for long-term in vivo use, which could have negative effects on its bioactivity . So a combination of collagen and PCL properly may compensate their shortcomings when applied for tissue-engineering. Our results showed that compared to pure PCL fibers, collagen/PCL fibers possessed better hydrophilicity and flexibility. Due to the addition of collagen, water contact angle of collagen/PCL meshes dramatically decreased which could facilitate cells to attach and spread. Meanwhile, PCL improved scaffold's load-bearing abilities, and together with collagen made this composite material possess greater Young's modulus and slowed its degradation rate without additional treatments. Good flexibility and proper degradation rate were also important factors for a tissue-engineered conduit to resist rearing and stretching forces and retain stable shape during nerve regeneration process . If nerve conduits had large rigidity, it might exert chronic compression to regenerated nerves. If too soft without enough mechanical support, it is impossible to be manipulated in an implant surgery and the conduits could not bear forces and collapse in vivo. Panseri S et al. found their electrospun PLGA/PCL tubular conduits collapsed in 40% of the treated rats for a 10 mm sciatic nerve gap repair . Therefore it is essential to explore the details on mechanical characteristics of nerve conduits. Our mechanical testing data and animal experiment results showed that the combination of collagen and PCL had good flexibility and gave good performance in resisting collapse and stretch forces in vivo which might provide some detail evidence for deepening our mechanical understanding to synthetic nerve conduits.
Different from central nervous system, peripheral nerves have some kind of capacity to regenerate after injuries. In this process, Schwann cell, the dominating glial cell type in peripheral nervous system plays crucial roles. They undergo changes of dedifferentiation and proliferation, secret neurotrophic factors and extracellular matrices and surround regenerating axons to form myelin sheaths which all contribute to a favorable environment for nerve regeneration . So before applying artificial nerve conduits in vivo, we first tested the biocompatibility of electrospun collagen/PCL fibrous material with Schwann cells through adhesion observation and proliferation assays in vitro. Previous research showed that Schwann cells well adhered on randomly aligned electrospun fibers, stretched across multiple fibers and elongated along the fiber axes . Our SEM observations were consistent with this finding. For proliferation assays, we found that postnatal Schwann cells could proliferate well just as on cover slips when seeded on electrospun fibers at the three observation time points. The above tests suggested that electrospun collagen/PCL composite material could be used as a good substrate for cell attachment and proliferation which agreed with previous results on the interactions of electrospun collagen/PCL fibers with glial cells [9, 23].
For in vivo test in peripheral nerve regeneration studies, sciatic nerve model is most commonly used for its adequate length and space at the mid-thigh for experimenters to manipulate a surgery and implant a graft easily . In our experiment, we used adult F344 rats to test the efficacy of electrospun collagen/PCL porous nerve conduits in bridging an 8 mm sciatic nerve defect gap. To evaluate axonal outgrowth and muscle reinnervation, a combination of traditional methods was used, including electrophysiology and histomorphometry.
For electrophysiological assessment, the amplitude of CMAP is one commonly used parameter which indirectly reflects the numbers of regenerated motor nerve fibers and the extent of muscle reinnervation while latency is an indirect parameter which refers to maturation of nerve fibers . Results of amplitude and latency measurements of CMAP in this experiment indicated collagen/PCL NCs group achieved similar functional muscle reinnervation as autografts. Besides, gastrocnemius muscle weight ratio is another parameter which could also conveniently give information about the efficacy of reinnervation by evaluating the extent of muscle atrophy in sciatic nerve lesion model . It is discovered that the maintenance of muscle mass is controlled by a balance between protein synthesis and protein degradation pathways. When a muscle is denervated as a consequence of nerve injury, the balance is destroyed and shift to degradation tendency which leads to decreased muscle cell size, muscle weight loss and hyperplasia of connective tissues . In this study compared to non-grafted conditions, electrospun collagen/PCL NCs successfully prevented serious muscle weight loss as autografts did either from muscle weight ratio comparisons or histological observations, which further necessitated nerve guidance structures in promoting nerve regeneration. It should be noted that although HE staining on paraffin sections is more susceptible to technical artifacts than on frozen ones, an apparent tendency of muscle atrophy prevention in collagen/PCL NCs group and autograft group can be clearly observed from their histological morphologies when compared to non-grafted group. Similar histological method and results could be found in related research [28, 29].
In nerve histology examination, regenerated nerves from autograft and collagen/PCL NCs groups all successfully grew through the gap and connected the proximal and distal nerve stumps. However, transmission electron microscopy and quantitative analyses of myelinated nerve fiber density and axonal diameter showed that those regenerated nerve fibers were still in a pre-mature stage of myelin sheath formation and maturation. Our experimental results of larger fiber density but relatively smaller-sized regenerated fibers in collagen/PCL NCs group might give evidence for the pruning hypothesis proposed by Brushart TM . After injury, proximal nerve stumps generated multiple collateral sprouts and over time nerve fibers failing to establish connections with target muscles or incorrectly projecting into the sensory branch would be pruned away . Therefore, we believe that on condition that observation time prolonged, nerve fiber density in collagen/PCL NCs group would decrease and nerve fibers would gradually become uniformly larger-sized and mature ones . In repair of nerve autograft, we speculate it is endoneurial tube that facilitates axonal outgrowth, which contributes to unnecessary sprouting following nerve injury, resulting in relatively fewer nerve fibers than in collagen/PCL NCs group.
Through a series of systematic sum and analysis researchers conclude that an appropriate combination of evaluation methods is preferred aiming at different research questions for different methods only illustrate discrete aspects of nerve regeneration [26, 33]. Because our exact aim is mainly to illustrate the role of nerve conduit in enhancing nerve fiber regeneration and the change of conduit itself in this process, functional or behavioral tests were not conducted in this experiment. For functional tests, walking track analysis is often adopted in rat sciatic nerve lesion model to assess motor functional recovery by calculating sciatic functional index (SFI) [[26, 34, 35]], while behavioral test such as Von-Frey test is used for sensory recovery measurement . Recently video recordings of rat gait have been developed to analyze motor function recovery in similar studies . These are all useful assessments to be applied in our future researches on functional recovery. Our examinations in this experiment demonstrated that although there were some morphological differences on regenerated nerve fibers, collagen/PCL NCs achieved similar electrophysiological and muscle histological results as autografts. As an artificial nerve conduit, electrospun collagen/PCL NC successfully protected regenerated nerves through the lesion defect from interruption of scar tissues, thus shortened time to earliest muscle reinnervation which was thought as the determinant in functional recovery .
Moreover, at the time of observation endpoint (4 months postoperatively), the implanted nerve conduits in our experiment almost degraded, losing their original guidance structures which well matched the nerve regeneration rate. Degradation rate is emphasized as an important design consideration for artificial nerve conduits because non-degradable tubes would compress regenerated nerves in the long term. Studies of functional recovery of transected peripheral nerves have indicated that there is a critical time period (the first 10-12 weeks following injury) for the regenerating axons to grow through the distal nerve stump for optimal recovery to occur [37, 38]. During this time period a relatively "undisturbed" microenvironment is required and afterwards a tendency of nerve conduit degradation is preferred. In vitro degradation test of our electrospun collagen/PCL scaffold showed that after 3-month immersion in PBS at a constant temperature of 37°C an approximate amount of 81.7% weight was lost as the scaffold degraded and became amorphous substances (data was not shown). Based on predecessors' work and our preliminary results, we finally chose a time point of 4 month to observe the conditions of nerve regeneration and conduit degradation in vivo.
Our macroscopic and microscopic results proved that the electrospun collagen/PCL nerve conduit with a wall thickness of 100-120 μm facilitated more axons regenerating through the gap and almost degraded at the observation endpoint of 4 months after injury (a time point shortly after the critical stage of axonal regeneration). Thus, we conclude that the degradation rate of electrospun collagen/PCL nerve conduits is well tailored to the need of a relatively "undisturbed" microenvironment for favorable peripheral nerve regeneration. The advantages of electrospun collagen/PCL nerve conduits on mechanical shape retaining and degradation rate controlling might give some useful supplements to previously published work, in particular those related researches on electrospun tubes in this field .